Nonlinear difference frequency processes, such as optical parametric oscillation, optical parametric amplification and optical parametric generation are used to generate longer wavelengths from a shorter wavelength via a process called difference frequency generation (DFG) using a nonlinear crystal. In optical parametric oscillation for example, the crystal converts the pump wavelength (i.e., the input wavelength) into two longer wavelengths—the signal and the idler. The nonlinear conversion generally requires high intensities of light, and, as such, the crystal may be placed inside a cavity to enhance the electric field of some or all of the wavelengths involved. The cavity can resonate one, two, or three wavelengths to increase the electric field of the light inside the cavity. A device in which light is resonated to convert longer wavelengths is called an optical parametric oscillator (OPO), an optical parametric amplifier (OPA), or an optical parametric generator (OPG).
Generally, the light-to-light conversion efficiency of an OPO is approximately one-half of the quantum limit. For example, if a 1 μm to a 4 μm conversion is desired, the expected efficiency would be about 0.5*1 μm/4 μm, or 12.5%. Reported values are typically around 10%, probably due to additional absorption in typical crystals used for this conversion. In order to generate 4 μm from 1 μm, seeds at 4 μm and 1.45 μm are required. If only the pump pulse is used, as is typical in an OPO, the seeds are quantum noise. It takes time to amplify the quantum noise, and, as such, this is one of the factors that contribute to the efficiency being lower than the quantum limit.
During this build up time, the pump is not efficiently converted to the signal and idler; the signal and idler are being amplified from noise to detectable levels. Thus, this pump power does not achieve significant conversion, and can be considered to be wasted. Typically, this build up time is a significant portion of the pump pulse duration. One way to increase efficiency might be to provide two sources: a pump and a signal or idler (seeder). However, seeding the nonlinear process adds a significant amount of complexity and cost since two sources must be used and synchronized. In addition, after a significant amount of signal and idler have been generated, efficiency may also be reduced due to back conversion. In this process, a signal and an idler photon are combined to produce a pump photon. For example, if a pulse with a temporal profile of a Gaussian is used, the build up time will be long since the pump intensity gradually increases, and back conversion will occur since the pump intensity is not constant. That is, the conversion efficiency will rise and fall as the pulse intensity changes. The conversion is only optimized for specific values of pump intensity.
Other approaches to improving OPO efficiency have included: 1) improvement of the crystal quality, minimizing absorption, and maximizing the nonlinear gain in the case of periodically poled materials; 2) use of an OPA after the OPO to convert a portion of the remaining pump light to the desired wavelength; 3) modification of the OPO cavity either by using multiple crystals or by optimization of the mirror reflectivities and curvatures; 4) design of the OPO to include multiple conversion processes to obtain the desired wavelength at higher powers.
Changing the OPO parameters will generally only lead to efficiencies of about half of the quantum limit. In addition, most of the parameters are fixed for a given system and cannot be changed in real time during OPO operation. If they can be changed, generally it is quite costly and time consuming process. Therefore the optimization is limited since many experimental variables are not known and cannot be accurately modeled. Inserting an OPA after the OPO increases the cost and complexity of the system. Using OPOs with multiple processes, e.g. an OPO with an additional crystal for difference frequency generation, requires that all processes be simultaneously phase matched, thereby dramatically increasing the sensitivity to temperature and manufacturing tolerances. In addition, additional material is added to the OPO which may cause additional absorption of the radiation.
There is thus a need for a system and method for increasing difference frequency generator efficiency while maintaining the architecture of the difference frequency generator.